Visual medical decision-making: Bipartite graphs vs. interactive tables.

Abstract

Most of the current medical diagnosis support systems are based on a textual design. In this thesis we present a model that uses a different design. It uses visualization to aid home diagnosis of common diseases in a user-friendly way. The model clearly displays the diagnostic results on the screen. A way of organizing the information into a picture of all symptoms, diseases, and the complex relationships between them (especially the combination of symptoms onto a single screen to give a global view) is presented. The purpose of designing this model is to bring complicated medical knowledge to the ordinary user. We believe that the simplified and economic display can demystify medicine, and empower the user to take better care of himself. By this convenient software tool people can discover quickly at home whether their symptom is serious or not, and then decide whether it is necessary to see the doctor; also people can compare the diagnosis the model makes with the doctors'. This model does not recommend treatment or therapy.